BestHub
Sign in
Tagged articles

LLM limitations

6 articles · Page 1 of 1
Machine Learning Algorithms & Natural Language Processing
Machine Learning Algorithms & Natural Language Processing
Oct 9, 2026 · Artificial Intelligence

Why LLMs Can't Count Letters in 'Strawberry': The Tokenization Problem Explained

This article explains how tokenization works in large language models, why it causes failures like miscounting letters in 'strawberry', details the BPE algorithm, discusses costs like language inequality and security risks, and explores alternatives like byte-level and dynamic tokenization.

BPELLM limitationsLarge Language Models
0 likes · 20 min read
Why LLMs Can't Count Letters in 'Strawberry': The Tokenization Problem Explained
AI Large-Model Wave and Transformation Guide
AI Large-Model Wave and Transformation Guide
Sep 11, 2026 · Artificial Intelligence

Palantir's Ontology Decoded: Six Mechanism Layers for Trustworthy Enterprise AI

This article dissects Palantir's ontology into six mechanism layers—semantic transparency, constrained query, controlled action, organizational unification, scenario generalization, and governed evolution—showing how each solves a specific AI deployment failure mode, why alternatives fall short at scale, and when the investment pays off.

AI deploymentEnterprise AILLM limitations
0 likes · 54 min read
Palantir's Ontology Decoded: Six Mechanism Layers for Trustworthy Enterprise AI
AI Large-Model Wave and Transformation Guide
AI Large-Model Wave and Transformation Guide
Sep 9, 2026 · Artificial Intelligence

Why LLMs Fail at Critical Decisions: Ontologies Provide the Missing World Model

The author details how LLMs failed in underwater battlefield simulations despite trying prompts, agents, RAG, and knowledge graphs, and explains why ontologies — formal, reasoning-capable world models — are essential for trustworthy AI decisions, illustrating with a concrete case and a six-step ontology engineering process.

AI decision-makingLLM limitationsOntoL
0 likes · 12 min read
Why LLMs Fail at Critical Decisions: Ontologies Provide the Missing World Model
AI Cyberspace
AI Cyberspace
Sep 8, 2026 · Artificial Intelligence

General Agent Harness Architecture: From 4 LLM Defects to 36 Functional Modules

This article derives a universal Agent Harness architecture from four mainstream coding agents, identifying four fundamental LLM limitations — no external perception, no action capability, no memory across calls, and no guaranteed correctness — and mapping them to eight responsibility domains with 36 concrete functional modules, plus a five-dimensional framework for vertical scenario adaptation.

Action ControlCoding AgentsHarness engineering
0 likes · 53 min read
General Agent Harness Architecture: From 4 LLM Defects to 36 Functional Modules
Thought Artisan
Thought Artisan
Mar 18, 2026 · Artificial Intelligence

Jeremy Howard: AI's Illusion of Software Engineering Competence

Jeremy Howard argues that AI tools only interpolate existing code patterns and cannot perform true software engineering, which requires understanding component interactions and system design; overreliance on AI erodes developer skills and creates unmaintainable technical debt, especially harming mid-career engineers.

AI codingJeremy HowardLLM limitations
0 likes · 8 min read
Jeremy Howard: AI's Illusion of Software Engineering Competence